We propose to develop methods and software tools for the statistical analysis of a new generation of gene expression microarrays. These new arrays are probe-rich in the sense that they provide multiple probes to measure the relative abundance of each and every known or predicted exon in the genome. Because of the large number and uniform placement of probes on the transcript, we can now obtain genome-wide survey of gene expression at the exon level. This will enable us to investigate new biological questions, such as alternative splicing, that were not assessable by previous generations of arrays. The result of this research will have great biological and medical significance as a large percentage of human genes are alternatively spliced and many diseases are linked to aberrations in splicing. We will use the Affymetrix Exon 1.0 ST arrays and next-generation Affymetrix Exon arrays as the motivating example for our research. We will develop methods for background noise corrections and for modeling cross-hybridization effects. Combining these with a probe-selection strategy, we will design a gene-level expression index for quantitative assessment of gene expression. Methods will be designed for the identification of new and /or alternatively spliced exons, and for the detection of differential splicing. Data and software generated in this research will be made freely available for public use. PUBLIC HEALTH RELEVANCE: This project will lead to new methods and software for the analysis of probe-rich microarrays. This will allow researchers to obtain better quantitative gene expression measurements and to study new biological questions such as alternative splicing. It is expected that many basic and translational biomedical studies critical for the improvement of public health will benefit from the results of this project. [unreadable] [unreadable] [unreadable]